Contents
Why gold reacts to data it should not care about
Gold does not respond to a number because of what the number says about gold. It responds because of what the number does to the dollar and to real interest rates, because those are the two variables that determine the opportunity cost of holding a non-yielding asset. Higher real rates make gold less attractive, and a stronger dollar makes gold more expensive for everyone else to buy.
The practical consequence is that the reaction is often counterintuitive. A strong jobs report, which sounds like good news, frequently sends gold down, because it pushes rate expectations up. Traders who trade the headline rather than the mechanism consistently get this backwards.
The specific chains for the main releases are in Fed decision trading and CPI and inflation trading.
The two-phase reaction
Almost every significant release produces two distinct phases, and they are separated by a pause of several minutes. The first phase is the headline reaction, which is fast, large and often partly a liquidity event, since market makers need to adjust to a new price and the first wave is priced on the number alone.
The second phase is the interpretation, and it is usually smaller and more persistent. This is where the market decides whether the detail under the headline justifies the move. A weak jobs number followed by a strong wage component can see gold rally on the headline and then reverse, because the wage data is what the Fed actually cares about.
The most useful trade in a news event is usually not the first spike. It is the second move, entered on a pullback after the initial reaction has settled. Traders who wait fifteen to thirty minutes and trade the interpretation of the pullback are trading with much better information than those who enter into the spike.
The specific releases and what they do
US employment data, the first Friday of the month, is the largest scheduled volatility event in gold. The headline non-farm payrolls number gets the initial reaction, and the unemployment rate and average hourly earnings components often drive the second phase. A payrolls beat with a soft wage growth figure is a very different setup from a payrolls beat with hot wages.
CPI, released around 18:30 IST, matters through the core component and the year-on-year figure. Headline CPI frequently disagrees with core, and the core reading is the one that moves rate expectations and therefore gold.
FOMC decisions are different in character. The announcement produces an immediate move that is frequently followed by an extended two-way market, because the statement and the dot plot are digested slowly. The press conference adds a second layer of volatility that lands roughly half an hour after the decision. We cover the full playbook in FOMC trading for gold.
Position sizing around events
The single most important rule is to reduce size before a scheduled release, not to increase it. Spreads widen during the release spike, which mechanically increases the cost of any trade you take, and the volatility increases the probability of a stop being hit by noise rather than by a genuine thesis failure.
A reasonable approach is to halve your normal risk on the release itself and to not hold a position through it unless the position was opened with the event in mind. If you do trade the event, size it so that the widened spread is a normal part of the cost calculation rather than a surprise.
Also be aware of the gap risk. If you hold a position over a weekend or a holiday with a tight stop, the release can open far beyond your stop, and you will be filled at the open price rather than at your stop. ATR-based stops cover sizing stops to volatility, which reduces but does not eliminate this.
The mistakes
The first is trading the headline interpretation rather than the number. If you have a view that “good news is good for gold”, you will be wrong more often than right in gold, because gold's reaction to good economic news is usually negative.
The second is entering into the initial spike. The spread at that moment is at its widest and the direction frequently reverses within minutes, which is exactly the condition where a stop-out is most likely and most expensive.
The third is trading every release. There are dozens of scheduled events, most of which produce noise rather than a tradable move. A trader who takes every one accumulates costs and noise. A trader who takes only the tier-one events, employment, CPI, and FOMC, has a far better chance of a good outcome.
The fourth is checking the calendar the night before and then forgetting. We keep a reference for this in the economic calendar guide, and news API access covers wiring it into a terminal.